Chapter 14

How to insert elements into ndarray

insert() ==> inserting the element at the required position.

append() ==> inserting the element at the end.

insert()

Example

Python

import numpy as np
help(np.insert)

Output

PowerShell

Help on function insert in module numpy:

insert(arr, obj, values, axis=None)
    Insert values along the given axis before the given indices.

Example

Python

# insert(array, obj, values, axis=None)
# Insert values along the given axis before the given indices.

#obj-->Object that defines the index or indices before which 'values' are inserted.
#values--->Values to insert into array.
#axis ---->Axis along which to insert 'values' .

Output

PowerShell

output nhi h

1-D arrays

Example

Python

# To insert 7777 before index 2
a = np.arange(10)
b = np.insert(a,2,7777)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [0    1 7777    2    3    4    5    6    7    8    9]

Example

Python

# To insert 7777 before index 2 and 8888 before index 5?
a = np.arange(10)
b = np.insert(a,[2,5],[7777,8888])
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b :[    0    1 7777    2    3    4 8888    5    6    7    8    9]

Example

Python

# To insert 7777 before index 2 and 8888 before index 5?
a = np.arange(10)
b = np.insert(a,[2,5],[7777,8888])
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [    0    1 7777    2    3    4 7777    5    6    7    8    9]

Example

Python

# shape mismatch
a = np.arange(10)
b = np.insert(a,[2,5],[7777,8888,9999])
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                              Traceback (most recent call last)
<ipython-input-358-735011b8a30e> in <module>
      1 # shape mismatch
      2 a = np.arange(10)
----> 3 b = np.insert(a,[2,5],[7777,8888,9999])
      4 print(f"array a : {a}")
      5 print(f"array b : {b}")
      
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
   4676      slobj[axis] = indices
   4677      slobj2[axis] = old_mask
-> 4678      new[tuple(slobj)] = values
   4679      new[tuple(slobj2)] = arr
   4680

ValueError: shape mismatch: value array of shape (3,)  could not be broadcast to indexing resultof shape (2,)

Example

Python

# shape mismatch
a = np.arange(10)
b = np.insert(a,[2,5,7],[7777,8888])
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                                    Traceback (most recent call last)
<ipython-input-359-d8cdd8c18d4d> in <module>
      1 # shape mismatch
      2 a = np.arange(10)
----> 3 b = np.insert(a,[2,5,7],[7777,8888])
      4 print(f"array a : {a}")
      5 print(f"array b : {b}")
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
   4676      slobj[axis] = indices
   4677      slobj2[axis] = old_mask
-> 4678      new[tuple(slobj)] = values
   4679      new[tuple(slobj2)] = arr
   4680
ValueError: shape mismatch: value array of shape (2,) could not be broadcast to indexing result of shape (3,).

Example

Python

a = np.arange(10)
b = np.insert(a,[2,5,5],[777,888,999])
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [ 0    1 777    2    3    4 888 999    5    6    7    8    9]

Example

Python

# IndexError
a = np.arange(10)
b = np.insert(a,25,7777)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
IndexError                              Traceback (most recent call last)
<ipython-input-361-423ff78357e9> in <module>
      1 # IndexError
      2 a = np.arange(10)
----> 3 b = np.insert(a,25,7777)
      4 print(f"array a : {a}")
      5 print(f"array b : {b}")
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
   4630      index = indices.item()
   4631      if index < -N or index > N:
-> 4632          raise IndexError(
   4633              "index %i is out of bounds for axis %i with "
   4634              "size %i" % (obj, axis, N))
IndexError: index 25 is out of bounds for axis 0 with size 10

Note

All the insertion points(indices) are identified at the beginning of the insert operation.

Array should contain only homogeneous elements.

By using insert() function, if we are trying to insert any other type element, then that element will be converted to array type automatically before insertion.

If the conversion is not possible then we will get error.

Example

Python

# the original array contains int values. If inserted value is float then the float value
# is converted to the array type i.e., int. There may be data loss in this scenario
a = np.arange(10)
b = np.insert(a,2,123.456)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [ 0    1 123    2    3     4     5     6     7     8    9]

Example

Python

a = np.arange(10)
b = np.insert(a,2,True)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [0 1 1 2 3 4 5 6 7 8 9]

Example

Python

a = np.arange(10)
b = np.insert(a,2,'GK')
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                             Traceback (most recent call last)
<ipython-input-364-57774f62be41> in <module>
      1 a = np.arange(10)
----> 2 b = np.insert(a,2,'GK')
      3 print(f"array a : {a}")
      4 print(f"array b : {b}")
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
  4638       # There are some object array corner cases here, but we canno
t avoid

   4639        # that:
-> 4640          values = array(values, copy=False, ndmin=arr.ndim, dtype=arr.dtype)
   4641          if indices.ndim == 0
   4642               # broadcasting is very different here, since a[:,0,:] = 
.. behaves

ValueError: invalid literal for int() with base 10: 'GK'

Example

Python

a = np.arange(10)
b = np.insert(a,2,10+20j)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
TypeError                               Traceback (most recent call last)
<ipython-input-365-d9aabac22b2c> in <module>
      1 a = np.arange(10)
----> 2 b = np.insert(a,2,10+20j)
      3 print(f"array a : {a}")
      4 print(f"array b : {b}")
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
   4638        # There are some object array corner cases here, but we cannot avoid
   4639        # that:
-> 4640      values = array(values, copy=False, ndmin=arr.ndim, dtype=arr.dtype)
4641         if indices.ndim == 0:
4642             # broadcasting is very different here, since a[:,0,:] = 
.. behaves

TypeError: can't convert complex to int

Summary for 1-D arrays while insertion

The number of indices and the number of elements should be matched.

Out of range index is not allowed.

Elements will be converted automatically to the array type.

2-D arrays

If we are trying to insert elements into multi dimensional arrays, compulsory we have to provide axis.

If we are not providing axis value, then default value None will be considered.

In this case, array will be flatten to 1-D array and then insertion will be happend.

axis=0 means rows (axis=-2).

axis=1 means columns (axis=-1).

Example

Python

# if the axis is not defined for 2-D arrays, None is selected by default
# here the 2-D array flatten to 1-D array and insertion will be happened
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,100)
print(f"array a :\n {a}")
print(f"array b : {b}")

Output

PowerShell

array a :
[[10 20]
[30 40]]
array b : [ 10  100  20  30  40]

figar banana

Example

Python

# insert the elements along the axis=0 or axis=-2 ==> rows are inserted
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,100,axis=0)
c = np.insert(a,1,100,axis=-2)
print(f"array a :\n {a}")
print(f"array b :\n {b}")
print(f"array c :\n {c}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[ 10 20]
 [100 100]
 [ 30 40]]
array c :
 [[ 10 20]
 [100 100]
 [ 30 40]]

Example

Python

# insert the elements along the axis=1 or axis=-1 ==> rows are inserted
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,100,axis=1)
c = np.insert(a,1,100,axis=-1)
print(f"array a :\n {a}")
print(f"array b :\n {b}")
print(f"array c :\n {c}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[ 10 100 20]
 [ 30 100 40]]
array c :
 [[ 10 100 20]
 [ 30 100 40]]

Example

Python

# to insert multiple rows
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,[[100,200],[300,400]],axis=0)

print(f"array a :\n {a}")
print(f"array b :\n {b}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[ 10 20]
 [100 200]
 [300 400]
 [ 30 40]]

Example

Python

# to insert multiple columns
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,[[100,200],[300,400]],axis=1)

print(f"array a :\n {a}")
print(f"array b :\n {b}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[ 10 100 300 20]
 [ 30 200 400 40]]

Example

Python

# ValueError
a = np.array([[10,20],[30,40]])
b = np.insert(a,1,[100,200,300],axis=0)

print(f"array a :\n {a}")
print(f"array b :\n {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                             Traceback (most recent call last)
<ipython-input-371-926c0e00dba4> in <module>
      1 # ValueError
      2 a = np.array([[10,20],[30,40]])
----> 3 b = np.insert(a,1,[100,200,300],axis=0)
      4
      5 print(f"array a :\n {a}")
      
<__array_function__ internals> in insert(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in insert(arr, obj, values, axis)
   4650        new[tuple(slobj)] = arr[tuple(slobj)]
   4651        slobj[axis] = slice(index, index+numnew)
-> 4652        new[tuple(slobj)] = values
   4653        slobj[axis] = slice(index+numnew, None)
   4654        slobj2 = [slice(None)] * ndim
   
ValueError: could not broadcast input array from shape (1,3) into shape (1,2)

append()

By using insert() function, we can insert elements at our required index position.

If we want to add elements always at end of the ndarray, then we have to go for append() function.

Example

Python

import numpy as np
help(np.append)

Output

PowerShell

Help on function append in module numpy:

append(arr, values, axis=None)
    Append values to the end of an array.

1-D arrays

If appended element type is not same type of array, then elements conversion will be happend such that all elements of same common type.

Example

Python

# append 100 element to the existing array of int
a = np.arange(10)
b = np.append(a,100)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [ 0  1  2  3  4  5  6  7  8  9 100]

Example

Python

# if the inserted element is not the type of the array then common type conversion
happened
a = np.arange(10)
b = np.append(a,20.5)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [ 0. 1.  2.  3.  4.  5.  6.  7.  8.  9.  20.5]

Example

Python

a = np.arange(10)
b = np.append(a,"GK")
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : ['0' '1' '2' '3' '4' '5' '6' '7' '8' '9' 'GK']

Example

Python

a = np.arange(10)
b = np.append(a,"GK")
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [0 1 2 3 4 5 6 7 8 9 1]

Example

Python

a = np.arange(10)
b = np.append(a,20+15j)
print(f"array a : {a}")
print(f"array b : {b}")

Output

PowerShell

array a : [0 1 2 3 4 5 6 7 8 9]
array b : [ 0. +0.j 1. +0.j 2. +0.j 3. +0.j 5. +0.j 5. +0.j 6. +0.j 7. +o.j 8. +0.j 9. +0.j 20.+15.j]
+0.j 8. +0.j 9. +0.j 20.+15.j]

2-D arrays

If we are not specifying axis, then input array will be flatten to 1-D array and then append will be performed.

If we are providing axis, then all the input arrays must have same number of dimensions, and same shape of provided axis.

if axis-0 is taken then the obj should match with the number of columns of the input array.

if axis-1 is taken then the obj should match with the number of rows of the input array.

Example

Python

# here axis is not specified so the default "None" will be taken and flatten to 1-D
array & insertion
a = np.array([[10,20],[30,40]])
b = np.append(a,70)
print(f"array a :\n {a}")
print(f"array b : {b}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b : [10 20 30 40 70]

Example

Python

# axis=0 ==> along rows
# Value Error : i/p array is 2-D and appended array is 0-D
a = np.array([[10,20],[30,40]])
b = np.append(a,70,axis=0) # appended is 0-D => 70
print(f"array a :\n {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                             Traceback (most recent call last)<ipython-input-379-c2af6bdb12eb> in<module>

      2 # Value Error : i/p array is 2-D and appended array is 0-D
      3 a = np.array([[10,20],[30,40]])
----> 4 b = np.append(a,70,axis=0) # appended is 0-D => 70
      5 print(f"array a :\n {a}")
      6 print(f"array b : {b}")
<__array_function__ internals> in append(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in append(arr, values, axis)
   4743           values = ravel(values)
   4744           axis = arr.ndim-1
-> 4745       return concatenate((arr, values), axis=axis)
   4746
   4747
<__array_function__ internals> in concatenate(*args, **kwargs)

ValueError: all the input arrays must have same number of dimensions, butthe array at index 0 has 2 dimension(s) and the array at index 1 has 0dimension(s)

Example

Python

# Value Error : i/p array is 2-D and appended array is 1-D
a = np.array([[10,20],[30,40]])
b = np.append(a,[70,80],axis=0) # appended is 1-D => [70,80]
print(f"array a :\n {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                             Traceback (most recent call last)
<ipython-input-380-d1f0cddf5145> in <module> 

      1 # Value Error : i/p array is 2-D and appended array is 1-D
      2 a = np.array([[10,20],[30,40]])
----> 3 b = np.append(a,[70,80],axis=0) # appended is 1-D => [70,80]
      4 print(f"array a :\n {a}")
      5 print(f"array b : {b}")
      
<__array_function__ internals> in append(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in append(arr, values, axis)
   4743         values = ravel(values)
   4744            axis = arr.ndim-1
-> 4745        return concatenate((arr, values), axis=axis)
   4746
   4747
   
<__array_function__ internals> in concatenate(*args, **kwargs)

ValueError: all the input arrays must have same number of dimensions, but the array at index 0 has 2 dimension(s) and the array at index 1 has dimension(s).

Example

Python

# i/p array is 2-D and appended array is 2-D
a = np.array([[10,20],[30,40]])
b = np.append(a,[[70,80]],axis=0) # appended is 2-D => [[70,80]]
# i/p array is 2-D and appended array is 2-D
a = np.array([[10,20],[30,40]])
b = np.append(a,[[70,80]],axis=0) # appended is 2-D => [[70,80]]

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[10 20]
 [30 40]
 [70 80]]

Example

Python

# axis=1 along colums
# Value Error : i/p array is Size 2 and appended element is size 1
a = np.array([[10,20],[30,40]])
b = np.append(a,[[70,80]],axis=1) # appended is size 1 => [[70,80]]
print(f"array a :\n {a}")
print(f"array b : {b}")

Output

PowerShell

---------------------------------------------------------------------------
ValueError                              Traceback (most recent call last)
<ipython-input-382-469d4a0f90ba> in <module>

      2 # Value Error : i/p array is Size 2 and appended element is size 1
      3 a = np.array([[10,20],[30,40]])
----> 4 b = np.append(a,[[70,80]],axis=1) # appended is size 1 => [[70,80]]
      5 print(f"array a :\n {a}")
      6 print(f"array b : {b}")
<__array_function__ internals> in append(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in append(arr, values, axis)
   4743         values = ravel(values)
   4744         axis = arr.ndim-1
-> 4745     return concatenate((arr, values), axis=axis)
   4746
   4747
   
<__array_function__ internals> in concatenate(*args, **kwargs)

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 2 and the array at index 1 has size 1

Example

Python

a = np.array([[10,20],[30,40]])
b = np.append(a,[[70],[80]],axis=1)
print(f"array a :\n {a}")
print(f"array b :\n {b}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[10 20 70]
 [30 40 80]]

Example

Python

# multiple columns
a = np.array([[10,20],[30,40]])
b = np.append(a,[[70,80],[90,100]],axis=1)
print(f"array a :\n {a}")
print(f"array b :\n {b}")

Output

PowerShell

array a :
 [[10 20]
 [30 40]]
array b :
 [[ 10 20 70 80]
 [ 30 40 90 100]]

Example

Python

# Consider the array?
a = np.arange(12).reshape(4,3)

# Which of the following operations will be performed successfully?
# A. np.append(a,[[10,20,30]],axis=0) #valid
# B. np.append(a,[[10,20,30]],axis=1) #invalid
# C. np.append(a,[[10],[20],[30]],axis=0) #invalid
# D. np.append(a,[[10],[20],[30],[40]],axis=1) #valid
# E. np.append(a,[[10,20,30],[40,50,60]],axis=0) #valid
# F. np.append(a,[[10,20],[30,40],[50,60],[70,80]],axis=1) #valid

Output

PowerShell

output nhi h

Example

Python

np.append(a,[[10,20,30]],axis=0) #valid

Output

PowerShell

array([[ 0, 1, 2],
       [ 3, 4, 5],
       [ 6, 7, 8],
       [ 9, 10, 11],
       [10, 20, 30]])

Example

Python

np.append(a,[[10,20,30]],axis=1) #invalid

Output

PowerShell

---------------------------------------------------------------------------
ValueError                             Traceback (most recent call last)
<ipython-input-387-4f71b3d245ce> in <module>
----> 1 np.append(a,[[10,20,30]],axis=1) #invalid

<__array_function__ internals> in append(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in append(arr, values, axis)
   4743             values = ravel(values)
   4744             axis = arr.ndim-1
-> 4745         return concatenate((arr, values), axis=axis)
   4746
   4747
   
<__array_function__ internals> in concatenate(*args, **kwargs)

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 0, the array at index 0 has size 4 and the array at index 1 has size 1.

Example

Python

np.append(a,[[10],[20],[30]],axis=0) #invalid.

Output

PowerShell

---------------------------------------------------------------------------
ValueError                              Traceback (most recent call last)
<ipython-input-388-52259d5553e8> in <module>
----> 1 np.append(a,[[10],[20],[30]],axis=0) #invalid

<__array_function__ internals> in append(*args, **kwargs)

F:\Users\Gopi\anaconda3\lib\site-packages\numpy\lib\function_base.py in append(arr, values, axis)
   4743            values = ravel(values)
   4744            axis = arr.ndim-1
-> 4745        return concatenate((arr, values), axis=axis)
   4746
   4747
   
<__array_function__ internals> in concatenate(*args, **kwargs)

ValueError: all the input array dimensions for the concatenation axis must match exactly, but along dimension 1, the array at index 0 has size 3 and the array at index 1 has size 1.

Example

Python

np.append(a,[[10],[20],[30],[40]],axis=1) #valid.

Output

PowerShell

array([[ 0, 1, 2, 10],
       [ 3, 4, 5, 20],
       [ 6, 7, 8, 30],
       [ 9,10,11, 40]])

Example

Python

np.append(a,[[10,20,30],[40,50,60]],axis=0) #valid.

Output

PowerShell

array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11],
       [10, 20, 30],
       [40, 50, 60]])

Example

Python

np.append(a,[[10,20],[30,40],[50,60],[70,80]],axis=1) #valid.

Output

PowerShell

array([[ 0,  1,  2, 10, 20],
       [ 3,  4,  5, 30, 40],
       [ 6,  7,  8, 50, 60],
       [ 9, 10, 11, 70, 80]])

Example

Python

a = np.arange(12).reshape(4,3)
print(a)
b = np.append(a,[[10],[20],[30],[40]],axis=1)
print(b)

Output

PowerShell

[[ 0  1  2]
 [ 3  4  5]
 [ 6  7  8]
 [ 9 10 11]]
[[ 0  1  2  10]
 [ 3  4  5  20]
 [ 6  7  8  30]
 [ 9 10 11 4 0]]

Example

Python

a = np.arange(12).reshape(4,3)
print(a)
b = np.append(a,[[10,50],[20,60],[30,70],[40,80]],axis=1)
print(b)

Output

PowerShell

[[ 0  1  2]
 [ 3  4  5]
 [ 6  7  8]
 [ 9 10 11]]
[[ 0  1  2  10  50]
 [ 3  4  5  20  60]
 [ 6  7  8  30  70]
 [ 9 10 11  40  80]

insert() vs append()

By using insert() function, we can insert elements at our required index position.

But by using append() function, we can add elements always at the end of ndarray.